社区博客与文章

Community Articles

TGI 多-LoRA：部署一次，搞定 30 个模型的推理服务

SmolLM：一个超快速、超高性能的小模型集合

NuminaMath 是如何荣膺首届 AIMO 进步奖的？

在 Hub 上使用 Presidio 进行自动 PII 检测实验

为视觉语言多模态模型进行偏好优化

在英特尔 Gaudi 2 上加速蛋白质语言模型 ProtST

Transformers 代码智能体成功刷榜 GAIA

Google 发布最新开放大语言模型 Gemma 2，现已登陆 Hugging Face Hub

微调 Florence-2 - 微软的尖端视觉语言模型

从 DeepSpeed 到 FSDP，再回到 Hugging Face Accelerate

欢迎 Stable Diffusion 3 加入 🧨 Diffusers

将强化学习重新引入 RLHF

英特尔 Gaudi 加速辅助生成

TGI 基准测试

NEW Articles from Team or Enterprise organizations will get promoted to the main section.

社区博客与文章

autotrust/JEV-27B-VL: a decision model that learned to see without a single image of training

Multimodal open d1 decision models for the edge

LightOnOCR-3: High-Performance OCR and Layout Extraction in One Model

Carbon-A: Finding genes in known and unknown genomes

Falcon OCR Arabic: 270M Parameters State-of-the-Art Arabic OCR

Falcon-Emirati: When an LLM Learns the Dialect, the Culture, and the Nuance

New in llama.cpp: Decision Models

ColGrep-Agent: Cutting Frontier Coding Agents’ Tokens by a Third

Darwin-27B-ZTC: A Single-Pass Judge and a Quantitative Look at Its Calibration

UIBK-Embedding-v1: A Dense and a Late-Interaction Retriever in One Model, One Index and One Search

The Open Quantum Challenge: Quantum Simulation and QEC Decoding on Classical GPUs

Sparseup-fr: A French replication of Linkup's sparse embedding model

NVIDIA Kumo Tabular Sets a New Accuracy-Efficiency Frontier for Tabular Prediction

Six Questions to Ask a LeRobot Dataset Before You Train

Uncensor any LLM with abliteration

Fine-Tune ASR for Your Domain with NVIDIA Speech Agent Skills

Leading the System One Mosaic Benchmark: What Darwin-27B-ZTC-v2's #1 Means

Basis Global Emotions: 525 hours of expressive speech across 108 locales

How to Use the Jev AI Model: A Step-by-Step Developer Guide

Open-sourcing AstaBrief, the fast report-generation model in Asta

TGI 多-LoRA：部署一次，搞定 30 个模型的推理服务

SmolLM：一个超快速、超高性能的小模型集合

NuminaMath 是如何荣膺首届 AIMO 进步奖的？

在 Hub 上使用 Presidio 进行自动 PII 检测实验

为视觉语言多模态模型进行偏好优化

在英特尔 Gaudi 2 上加速蛋白质语言模型 ProtST

Transformers 代码智能体成功刷榜 GAIA

Google 发布最新开放大语言模型 Gemma 2，现已登陆 Hugging Face Hub

微调 Florence-2 - 微软的尖端视觉语言模型

从 DeepSpeed 到 FSDP，再回到 Hugging Face Accelerate

欢迎 Stable Diffusion 3 加入 🧨 Diffusers

将强化学习重新引入 RLHF

英特尔 Gaudi 加速辅助生成

TGI 基准测试

autotrust/JEV-27B-VL: a decision model that learned to see without a single image of training

Multimodal open d1 decision models for the edge

LightOnOCR-3: High-Performance OCR and Layout Extraction in One Model

Carbon-A: Finding genes in known and unknown genomes

Falcon OCR Arabic: 270M Parameters State-of-the-Art Arabic OCR

Falcon-Emirati: When an LLM Learns the Dialect, the Culture, and the Nuance

New in llama.cpp: Decision Models

ColGrep-Agent: Cutting Frontier Coding Agents’ Tokens by a Third

Darwin-27B-ZTC: A Single-Pass Judge and a Quantitative Look at Its Calibration

UIBK-Embedding-v1: A Dense and a Late-Interaction Retriever in One Model, One Index and One Search

The Open Quantum Challenge: Quantum Simulation and QEC Decoding on Classical GPUs

Sparseup-fr: A French replication of Linkup's sparse embedding model

NVIDIA Kumo Tabular Sets a New Accuracy-Efficiency Frontier for Tabular Prediction

Six Questions to Ask a LeRobot Dataset Before You Train

Uncensor any LLM with abliteration

Fine-Tune ASR for Your Domain with NVIDIA Speech Agent Skills

Leading the System One Mosaic Benchmark: What Darwin-27B-ZTC-v2's #1 Means

Basis Global Emotions: 525 hours of expressive speech across 108 locales

How to Use the Jev AI Model: A Step-by-Step Developer Guide

Open-sourcing AstaBrief, the fast report-generation model in Asta